New 4.1 crash when calling rad scheme

All things involving the setup (compiling, NetCDF, etc.) of the WRF Model itself.

New 4.1 crash when calling rad scheme

Postby jcanellas » Tue Jun 18, 2019 9:57 am

Dear all,

Last week, I installed new 4.1 version of WRF, along with 4.0 version of WPS, in order to try it's new features. I was previously running WRF version 3.9, without major problems in the build. I compiled 4.1 using the same libraries, to wit, those recommended in the WRF compilation tutorial: http://www2.mmm.ucar.edu/wrf/OnLineTuto ... .php#STEP2. I compiled ARW with parallel support and the gnu fortran compiler (option 34 of configure file). I chose the em_real case.
When I tried mi first run, the following error occurred :



Timing for main: time 2017-12-13_06:01:20 on domain 1: 7.14854 elapsed seconds
d01 2017-12-13_06:01:20 grid spacing, dt, time_step_sound= 3000.00000 80.0000000 16
d01 2017-12-13_06:01:20 calling inc/HALO_EM_MOIST_OLD_E_7_inline.inc
d01 2017-12-13_06:01:20 calling inc/PERIOD_BDY_EM_MOIST_OLD_inline.inc
d01 2017-12-13_06:01:20 calling inc/HALO_EM_A_inline.inc
d01 2017-12-13_06:01:20 calling inc/PERIOD_BDY_EM_A_inline.inc
d01 2017-12-13_06:01:20 calling inc/HALO_EM_PHYS_A_inline.inc
d01 2017-12-13_06:01:20 Top of Radiation Driver
d01 2017-12-13_06:01:20 CALL cldfra1

(these are the final lines of the rls.out.0000 file)

The wrf.exe will stop, and it says:

Segmentation fault


As I was trying the newly incorporated Cheng scheme for shallow cumulus convection, I thought that this error would be somehow related to that scheme. Nevertheless, I repeated the run (rerun real.exe, then rerun wrf.exe) changing the shcu_physics options, and the same error happened again. I tried several combinations of physics options, and all of them returned this error. I thought maybe there was some relevant variable in the RegistriEM_COMMON file which was not properly declared when compiling, I looked for "cldfra" there, and I found several variables, none of which seemed to be scheme independent (I mean, the cloud fraction options I saw there where defined for some particular scheme, so I think they are not to blame for this error). Yet, I didn't recompiled, nor did I change anything in the Registry files.

I'm inclined to think this is somehow related to the WRF 4.1 compilation process, since I run the same model, version 3.9, using the same libraries directory, and everything works fine. I will attach, anyway, the namelist.input file I used, only remember that I changed several times all the physics options (mp_physics ra_lw_phys ra_sw_physics sf_sfclay_physics sf_surface_physics bl_pbl_physics cu_physics shcu_physics)

I'm a fresh PhD student, I started working with WRF only two months ago, and many of the options of the namelist I merely copied from someone else's file. I don't completely understand what they mean, but I've managed to make a few runs with this same file in the 3.9 version.

Thank you all in advance, anything you can say to point me in the right direction I will deeply appreciate.


 &time_control
run_days = 3,
run_hours = 18,
run_minutes = 0,
run_seconds = 0,
start_year = 2017, 2017, 2017,
start_month = 12, 12, 12,
start_day = 13, 11, 11,
start_hour = 06, 06, 06,
end_year = 2017, 2017, 2017,
end_month = 12, 12, 12,
end_day = 17, 19, 19,
end_hour = 00, 00, 00,
interval_seconds = 21600,
input_from_file = .true.,.true.,.true.,
history_interval = 180,60,12,
frames_per_outfile = 1000,1000,1000,
restart = .false.,
restart_interval = 5000,
io_form_history = 2
io_form_restart = 2
io_form_input = 2
io_form_boundary = 2
debug_level = 50



/

&domains
time_step = 80,
time_step_fract_num = 0,
time_step_fract_den = 1,
max_dom = 1,
s_we = 1,1,1,
e_we = 160, 118, 100,
s_sn = 1,1,1,
e_sn = 242, 181, 100,
s_vert = 1,1,1,
e_vert = 33,33,33,
p_top_requested = 5000,
num_metgrid_levels = 32,
num_metgrid_soil_levels = 4,
dx = 3000, 10000.0, 3333.3333333333335,
dy = 3000, 10000.0, 3333.3333333333335,
grid_id = 1,2,3,
parent_id = 1,1,2,
i_parent_start = 1, 20, 45,
j_parent_start = 1, 30, 70,
parent_grid_ratio = 1, 3, 3,
parent_time_step_ratio = 1,3,3,
feedback = 0,
smooth_option = 0,
/

&physics



mp_physics = 1, 28, 28,
ra_lw_physics = 1, 4, 4,
ra_sw_physics = 1, 4, 4,
sf_sfclay_physics = 1, 1, 1,
sf_surface_physics = 0, 1, 1,
bl_pbl_physics = 1, 5, 5,
cu_physics = 0, 0, 0,
shcu_physics = 0,
surface_input_source = 1,
num_soil_layers = 4,
sf_urban_physics = 0, 0, 0,
windfarm_opt = 0, 0, 0,
windfarm_ij = 0,
maxiens = 1,
maxens = 3,
maxens2 = 3,
maxens3 = 16,
ensdim = 144
slope_rad = 1, 1, 1,
topo_shading = 1, 1, 1,
shadlen = 25000,
bl_mynn_edmf = 0
num_land_cat = 21

/


&dynamics
w_damping = 0,
diff_opt = 1,
km_opt = 4,
diff_6th_opt = 0, 0, 0,
diff_6th_factor = 0.12, 0.12, 0.12,
base_temp = 290.,
damp_opt = 0,
zdamp = 5000., 5000., 5000.,
dampcoef = 0.2, 0.2, 0.2,
khdif = 0, 0, 0,
kvdif = 0, 0, 0,
non_hydrostatic = .true., .true., .true.,
moist_adv_opt = 1, 1, 1,
scalar_adv_opt = 1, 1, 1,
gwd_opt = 0,
/

&bdy_control
spec_bdy_width = 5,
spec_zone = 1,
relax_zone = 4,
specified = .true., .false., .false.,
nested = .false., .true., .true.,
/

&grib2
/

&namelist_quilt
nio_tasks_per_group = 0,
nio_groups = 1,
/
jcanellas
 
Posts: 8
Joined: Tue Jun 18, 2019 9:21 am
Location: Buenos Aires, Argentina

Re: New 4.1 crash when calling rad scheme

Postby jcanellas » Fri Jul 12, 2019 12:50 pm

Dear all,

It seems I've solved this issue, although I could not say how exactly. The main modification I did to the above namelist was to dramatically change the &domains section.
I'm sorry I can't give more precise information, but I hope this will help someone! :mrgreen:
jcanellas
 
Posts: 8
Joined: Tue Jun 18, 2019 9:21 am
Location: Buenos Aires, Argentina


Return to Installation

Who is online

Users browsing this forum: No registered users and 6 guests